Publication number | US7974833 B2 |
Publication type | Grant |
Application number | US 11/158,897 |
Publication date | Jul 5, 2011 |
Filing date | Jun 21, 2005 |
Priority date | Jun 21, 2005 |
Fee status | Paid |
Also published as | US20070016400 |
Publication number | 11158897, 158897, US 7974833 B2, US 7974833B2, US-B2-7974833, US7974833 B2, US7974833B2 |
Inventors | Radu Soricut, Daniel Marcu |
Original Assignee | Language Weaver, Inc. |
Export Citation | BiBTeX, EndNote, RefMan |
Patent Citations (60), Non-Patent Citations (144), Referenced by (32), Classifications (7), Legal Events (3) | |
External Links: USPTO, USPTO Assignment, Espacenet | |
Text to text applications include machine translation, automated summarization, question answering, and other similar applications where a machine carries out the function of understanding some kind of input information and generating text. The input information is often “text”, but can be any kind of information that is received and understandable by the machine.
Conventional text to text applications use heterogeneous methods for implementing the generation phase. Machine translation often produces sentences using application-specific decoders that are based on work that was conducted on speech recognition. Automated summarization produces abstracts using task specific strategies.
Text to text applications have struggled with use of generic natural language generation (NLG) systems, because they typically do not have access to the kind of information required by the formalisms of natural language generation systems. For example, natural language generation may require formalisms such as semantic representations, syntactic relations, and lexical dependencies. The formalisms also require information that is obtained from deep semantic relations, as well as shallow semantic relations and lexical dependency relation. Machine systems typically do not have access to deep subject verb or verb object relations.
A number of natural language generation systems are known, including FUF, Nitrogen, HALogen, and Fergus.
The formal language of IDL (Interleave; Disjunction; Lock) was proposed by Mark-Jan Nederhof and Giorgio Satta in their 2004 paper: IDLexpressions: a formalism for representing and parsing finite languages in natural language processing. Journal of Artificial Intelligence Research, 21: 287-317. Using IDL expressions, one can compactly represent word- and phrase-based encoded meanings. Nederhof and Satta also present algorithms for intersecting IDL expressions with non-probabilistic context free grammars.
A new language for compactly representing large sets of weighted strings is described. The original IDL language of Nederhof and Satta is extended to a weighted form that can be given a probabilistic interpretation. This language is called Weighted IDL (WIDL).
An aspect provides efficient algorithms for intersecting WIDL expressions with ngram and syntax-based language models. It thus enables one to create from the set of strings that are compactly represented by a WIDL expression those that are grammatical according to some external knowledge resources using those ngram- and syntax-based language models.
An aspect describes how WIDL expressions and the above mentioned intersection algorithms can be used in text-to-text natural language applications, such as machine translation and summarization.
According to one aspect of the present system, probability distributions are associated with weighted IDL (“WIDL) operators, to allow weighted IDL expressions to probabilistically represent biases over the entire set of strings that are subsumed by the weighted IDL expression.
The WIDL expressions may be intersected with various language model combinations, while preserving the compactness property of WIDL expressions. The output is a string that is encoded by the input WIDL expression that receives the highest score based on the combination of WIDL and language model scores.
These and other aspects will now be described in detail with reference to the accompanying drawings, in which:
Alternatively, the component that generates a WIDL expression can be used in conjunction with other text-to-text applications. For example, in summarization as shown in
By intersecting this WIDL expression with ngram and syntax-based language models via the same processor 100, one can generate in this instance sentences that are grammatical and likely to contain important information in a document. The same process can be used similarly in any other applications, such as speech recognition, paraphrasing, natural language generation, question answering, or the like—in any application in which one wishes to generate text.
The WIDL processor 100 may be any general purpose computer, and can be effected by a microprocessor, a digital signal processor, or any other processing device that is capable of executing the operations described herein.
The flowcharts described herein can be instructions which are embodied on a machine-readable medium such as a magnetic disc, optical disk, or any other medium that can be read by a computer. Alternatively, the flowchart can be executed by dedicated hardware, or by any known or later discovered processing device.
The present application describes use of weighted IDL operators. These operators use information that is available in many text to text applications to generate available information relevant to those text to text applications. This may include bias for word choice, word order, phrasal combination and the like.
Weighted IDL, as described herein, encodes meaning via words and phrases that are combined using a set of operator types. The operator types may include:
The language models 107 can be n-gram and syntax based language models, as shown. In one implementation of an ngram language model, the probability of a string is computed by multiplying the probabilities of each word in the string given n words before it. In one implementation of a syntax-based language model, one computes the probability of a hierarchical structure associated with a string (a syntactic tree) by multiplying the probabilities of creating each of the nodes in the hierarchical structure.
WIDL allows the “language”, which here refers only to the large set of strings being described by the WIDL, to be described using a very compact representation. Input to the model is produced, and the output of the generation process is the string that is encoded by the input expression that has received the highest score based on the combination of WIDL and language model scores.
IDL was described by Nederhoff and Sata, in their 2004 paper entitled “IDL-expressions: a formalism for representing and parsing finite languages in natural langauge processing” Journal of artificial intelligence research, 21:287-317. IDL expressions may be used as an application independent, information-slim representation language for natural language creation. IDL operators handle natural constraints such as word choice and precedents, constructions such as phrasal combination and under-specifications such as word order.
A brief description of IDL as it applies to this embodiment is provided. The concatenation (•)operator uses strings encoded by its argument expressions to obtain concatenated strings that respect the order of the arguments. For example, the concatenation a•b encodes the singleton set {ab}.
The interleave ∥operator interleaves strings encoded by its argument expression. For example, ∥(a.b,c) encodes the sets {cab, acb, abc}
The disjunction operator V allows a choice among the strings that are encoded by its argument expressions. For example, V(a,b) encodes the set {a,b}
The lock operator X takes only one argument, and locks in the strings encoded by its argument expression such that no additional material can be interleaved. For example, ∥(x (a•b), c) encodes the set {cab,abc}.
Consider the following IDL expression:
∥(finally, V(x(the•prisoners), x(the•captives))•were•released) (1)
The operators define the way the natural language generation works. The concatenation operator captures precedence constraints. For example, in expression 1, the determiner “the” appears before the noun it determines, here “prisoners and captives”. The lock operator X enforces the phrase encoding constraints. Here, the lock operator maintains the phrase “the captives” as a phrase which needs to be used as a whole, as well as maintaining the phrase “the captives”.
The disjunction operator V allows for multiple word phrase choices. That is, the disjunction between “the prisoners” and “the captives” allows either one to be used. The ∥operator allows for word order freedom. That is, word orders can be underspecified at meanings at the representation level. IDL allows many different strings to be encoded in this way, using a very compact representation. For example, expression 1 can encode the strings:
However, the following strings are not part of the language defined by expression 1:
the finally captives were released (this string is disallowed because the lock operator locks the phrase “the captives”)
the prisoners were released (this string is disallowed because the Interleave operator requires all of its arguments to be represented, and this phrase ignores the “finally”)
finally the captives released were (this string violates the order that is imposed by the precedence operator (were • released).
While the above has described IDL expressions in a text form, IDL graphs may be used as an equivalent method of representing the IDL expressions.
The graph G(Π) defines a finite language as the set of strings that can be generated from an IDL specific traversal of G(Π) resulting in this traversing the start to the end.
The IDL graphs can also be intersected with finite state acceptors. An advantage of the IDL representation is its compactness; the IDL need not enumerate all possible interleavings, the notation inherently represents these interleavings without stating them all expressly.
WIDL expressions are created from strings using the word choice/disjunction operator V, the concatenation or precedence operator (•), and the locked interleaved operator ∥. The WIDL expressions described herein have probability distributions associated with each V and ∥ operator. These probability distributions are used to induce a probability value to each of the strings encoded under the operators.
A formal definition of the alphabet of WIDL expressions follows
Definition 1 Let Σ be some finite alphabet (called | |
input alphabet), let Δ be another, disjoint, finite alphabet | |
(called distribution alphabet), and let ε be a | |
symbol not in Σ. A WIDL-expression over Σ is a | |
string π satisfying one of the following conditions: | |
1. | π = a, with a ∈ Σ ∪ {ε}; |
2. | π = V_{δ}(π_{1}, π_{2}, . . . ,π_{n}), with δ ∈ Δ, n ≧ 2, |
and π_{i }a WIDL-expression, 1 ≦ i ≦ n; | |
3. | π = ||_{δ}(π_{1}, π_{2}, . . . ,π_{n}), with δ ∈ Δ, n ≧ 2, |
and π_{i }a WIDL-expression, 1 ≦ i ≦ n; | |
4. | π = π_{1 }· π_{2}, with π_{1 }and π_{2 }WIDL-expressions |
The WIDL operators operate as follows: the concatenation operator (•) uses the strings encoded by its arguments to obtain concatenated strings in the order stated.
The V_{δ0 }operator defines disjunction with associated probability distribution δ_{0}, and allows for a choice among the strings encoded by its argument expressions. The associated probability distribution δ0 is used to induce a probability distribution δ over all the strings. For example, V_{δ0 }(a,b), δ0={1:0.8,2:0.2}, encodes the set {a,b}, over which the probability distribution δ is δ(a)=δ0(1)=0.8; δ(b)=δ0(2)=0.2.
The ∥_{δ0 }operator represents lock interleaved with associated probability distribution δ0. This interleaves the strings encoded by its argument expressions. This treats the strings of each argument expression as locked expressions, for example atomic expressions. For example, ∥δ0(a b,c), δ0={1 2:0.95, 2 1 ″0.05} encodes the set {abc,cab}, over which δ assigns the probabilitys δ(abc)=δ0(1 2)−0.95, δ(cab)−δ0 (2 1)=0.05.
Each WIDL expression over Σ and Δ and Υ has a formal interpretation as a set of strings from Σ and a probability distribution over that set of strings. The definition of Σ can be given as follows:
Definition 2 Let Σ be some finite alphabet (called | |
input alphabet), let Δ be another, disjoint, finite alphabet | |
(called distribution alphabet), and let σ be a | |
function mapping an WIDL-expression over Σ and | |
Δ into a set of strings in Σ^{+ }characterized by a probability | |
distribution δ ∈ Δ, defined as follows: | |
1. | σ(a) = {a},δ(a) = 1,for a ∈ Σ; |
2. | σ(ξ) = {∈},δ(∈) = 1; |
3. | σ(π_{1 }· π_{2}) = {e_{1}e_{2}|e_{1 }∈ σ(π_{1}),e_{2 }∈ |
σ(π_{2})},δ(e_{1}e_{2}) = δ_{1}(e_{1})·δ_{2}(e_{2}), where δ_{1 }and δ_{2} | |
are the probability distributions associated with σ(π_{1}) and σ(π_{2}), | |
respectively; | |
4. | σ(V_{δ} _{ 0 }(π_{1}, . . . ,π_{n})) = ∪_{i=1} ^{n}{e_{i}|e_{i }∈σ(π_{1})},δ(e_{i}) = p_{i }· δ_{i}(e_{i}), |
where δ_{0 }is specified over {1,...,n} as δ_{0}(i) = p_{i}, Σ_{i=1} ^{n}p_{i }= 1, | |
and δ_{i}s are the probability distributions associated with each σ(π_{i}), | |
1≦ i ≦ n; | |
5. | σ(||_{δ} _{ 0 }(π_{1}, . . . ,π_{n})) = ∪_{{i} _{ 1 } _{...i} _{ n } _{}∈Perm(n) }{e_{i} _{ 1 }. . . e_{i} _{ n }|e_{i} _{ k } ∈ σ(π_{i} _{ k })}; |
δ(e_{i} _{ 1 }...e_{i} _{ n }) = p_{k ·} δ_{i} _{ 1 }(e_{i} _{ 1 }) . . . δ_{i} _{ n }(e_{i} _{ n }), | |
where i_{1}...i_{n }∈ Perm(n) is the k-th permutation | |
of n elements^{1}, and δ_{0 }is specified over Perm(n) as δ_{0}(i_{1 }. . . i_{n}) = | |
p_{k}, Σ_{k=1} ^{n!}p_{k }= 1, and δ_{ik}s are the probability distributions associated with | |
each σ(π_{i} _{ k }), 1 ≦ i_{k }≦ n. | |
The ∥δ0 in WIDL is a weaker version of IDL's parallel operator, since the lock is probabilistic, rather than absolute.
Consider the following WIDL expression (2) corresponding to the above IDL expression (1)
∥_{δ} _{ 1 }(finally, V_{δ} _{ 2 }(the • prisoners, the • captives), were, released)
δ_{1}={1 2 3 4:0.25, 2 3 1 4:0.53, all other: 0.10}
δ_{2}={1:0.7,2:0.3} (2)
The language defined by the WIDL expression is weighted in the sense that it has a probability distribution associated with it. Each string in this language has an associated nonnegative real number less than or equal to one. The sum of the probabilities over the entire “language” adds up to one. The following strings, and their associated probabilities, are among those that are encoded by WIDL expression (2):
This string encodes the sequence 1234 (probability 0.25) for the probability of ∥δ1. It also encodes choice 1 for the argument of Vδ2 (probability 0.7) Therefore, it receives a probability of 0.25 times 0.7 equals 0.175.
The captives finally were released; probability 0.1 times 0.3 equals 0.03
This string encodes the sequence 2134 for the arguments of ∥δ0, (probability of 0.1) and choice two for the arguments of Vδ2. (Probability of 0.3.)
The prisoners were finally released; probability 0.53 times 0.7 equals 0.371.
This encodes the sequence 2314 for the arguments of ∥, and choice one for the arguments of Vδ2, probability is 0.53 times 0.7 equals 0.371.
WIDL expressions can be expressed in terms of an equivalent WIDL graph. A WIDL graph is denoted by the parameter set or “tuple” (V, E vs, ve, λ, r), where
V and E are finite sets of vertices and edges respectively,
vs and ve are special vertices in V called start vertices and end vertices respectively.
λ is the edge labeling function that maps the edge set E. into the alphabet ΣU {(ε, <δ,> δ, (δ,) δ}
r is the vertex ranking function, mapping V to the set of non-negative integer numbers.
The label ε indicates that an edge does not consume an input signal. The labels <δ,> δ, and (δ,)δindicate the same as ε, but also indicate that it is the start or end of something that corresponds to a ∥δ or a Vδoperator respectively. This information may be used to handle locked strings with the intended interpretation of the locked interleaved type of operator.
The following is a formalized definition of how a WIDL expression is mapped into its WIDL graph.
Each weighted IDL graph has an associate finite weighted language formed of a finite set of strings with associated probability distribution, as its set of strings. An associated probability that can be obtained from a weighted IDL specific traversal of the graph starts from the start vertex and ends at the end vertex. Therefore, a WIDL expression Π and its corresponding WIDL graphs Υ(Π) are said to be equivalent, since they generate the same finite weighted language, denoted as Lδ(Π).
The following text describes how to combine/intersect WIDL graphs with ngram and syntax-based language models by
An embodiment describes how to combine the IDL-graphs and Finite-State Acceptors. The IDL formalism is linked with the more classical formalism of finite-state acceptors (FSA) (Hopcroft and Ullman, 1979). The FSA representation can naturally encode precedence multiple choice, but it lacks primitives corresponding to the interleave (II) and lock (x) operators. As such, an FSA representation must explicitly enumerate all possible interleavings, which are inherently captured in an IDL representation. This correspondence between implicit and explicit interleavings is handled by the notion of a cut of an IDL-graph G (Π).
The IDL graph shown in
Intuitively, a cut through G(Π)is a set of vertices that can be reached simultaneously when traversing G from the initial node to the final node, following the branches as prescribed by the encoded I, D, and L operators, in an attempt to produce a string in L(Π). More precisely, the initial vertex v8 is considered a cut (
1) the vertex that is the start of several edges is labeled using the special symbol
. This is replaced by a sequence of all the end vertices of these edges (for example, v0v2 is a cut derived from v8, (A mirror rule handles the special symbol which is actually the end of several edges
.2) The vertex that is the start of an edge labeled using vocabulary items or ε is replaced by the end vertex of that edge (for example, v1v2, v0v3, v0v5, v0v6 are cuts derived from v0v2, v0v2, v0v3 and v0v5, respectively as shown in
Note that the last part of the second rule restricts set of cuts by using the ranking mechanism. If v1v6 could be a cut, it would imply that “finally” may appear inserted between the words of the locked phrase “the prisoners”.
The IDL formalism is linked with the FSA formalism by providing a mapping from an IDL-graph G(Π) to an acyclic finite-state acceptor A(Π). Because both formalisms are used for representing finite languages, they have comparable representational power.
The following denotes the state of A(Π) by the name of the cut to which it corresponds. A transition labeled α in A(Π) between states [vi′, . . . vk′ . . . v′j] and state [v″I . . . v″k . . . v″j] occurs if there is an edge (v′k, α, v″k in G(Π). For the example in
The conversion from the IDL representation to the FSA representation destroys the compactness property of the IDL formalism, because of the explicit enumeration of all possible interleavings, which causes certain labels to appear repeatedly in transitions. For example, a transition labeled finally appears 11 times in the finite state acceptor in
The Computational Properties of IDL-expressions are discussed herein. A first embodiment describes IDL-graphs and Weighted Finite-State acceptors.
As mentioned above, the generation mechanism that concerns this disclosure performs an efficient intersection of IDL expressions with n-gram language models. Following (Mohri et al., 2002; Knight and Graehl, 1998), language models are first implemented using weighted finite-state acceptors (wFSA). The above describes a mapping from an IDL-graph G(Π) to a finite-state acceptor A(Π). This finite-state acceptor A(Π), provides a weighted finite-state acceptor W(Π), by splitting the states of A(Π) according to the information needed by the language model to assign weights to transitions. For example, under a bigram language model LM, state [v1v16] in
This point therefore already represents intersecting IDL-expressions with n-gram language models. From an IDL-expression r, following the mapping Π->g(Π) ->A(Π)->W(Π), a weighted finite-state acceptor is obtained. This allows using a single-source shortest path algorithm for directed acyclic graphs (Cormen et al., 2001) to extract the realization corresponding to the most probable path.
Premature unfolding of the IDL-graph into a finite-state acceptor can destroy the representation compactness of the IDL representation. For this reason, techniques are dislosed that, although similar in spirit with the single-source shortest-path algorithm for directed acyclic graphs, perform on-the-fly unfolding of the IDL-graph, with a mechanism to control the unfolding based on the scores of the paths already unfolded. Such an approach has the advantage that prefixes that are extremely unlikely under the language model may be regarded as not so promising, and parts of the IDL expression that contain them may not be unfolded. This may lead to savings.
A disclosed technique describes generation via Intersection of IDL-expressions with Language Models Algorithm IDL-NGLM-BFS. This intersects an IDL-graph with an n-gram language model LM using incremental unfolding and breadth-first search.
The pseudocode for the IDL-NGLM-BFS is shown below:
IDL-NGLM-BFS(G, LM) | |
1 active {[vs^{G}]} | |
2 flag 1 | |
3 while flag | |
4 do unfold UNFOLDIDLG(active, G) | |
5 EVALUATENGLM(unfold, LM) | |
6 if FINALIDLG(unfold, G) | |
7 then flag 0 | |
8 active unfold | |
9 return active | |
This builds a weighted finite state acceptor W corresponding to an IDL-graph G incrementally, by keeping track of a set of active states, called active. The incrementality comes from creating new transitions and states in W originating in these active states, by unfolding the IDL graph G; the set of newly unfolded states is called unfold. The new transitions in W are weighted according to the language model.
If a final state of W is not yet reached, the while loop is closed by making the unfold set of states to be the next set of active states. Note that this is actually a breadth first search (BFS) with incremental unfolding. This technique still unfolds the IDL-graph completely, and therefore may have the potential drawback of lack of compactness.
An interesting contribution of algorithm IDL-NGLM-BFS, however, is its incremental unfolding. If, instead of line 8 in mechanisms to control which unfold states become part of the active state set for the unfolding iteration, additional techniques are obtained.
Technique IDL-NGLM-A* is obtained by modifying line 8, thus obtaining the pseudocode:
IDL-NGLM-A*(G, LM) | |
1 active {[vs^{G}]} | |
2 flag 1 | |
3 while flag | |
4 do unfold UNFOLDIDLG(active, G) | |
5 EVALUATENGLM(unfold, LM) | |
6 if FINALIDLG(unfold, G) | |
7 then flag 0 | |
8 for each state in unfold | |
do PUSH(astarQ, state) | |
active POP(astarQ) | |
9 return active | |
This uses a priority queue, astarQ, as a control mechanism from which the states from unfold are PUSH-ed, sorted according to an admissible heuristic function (Russell and Norvig, 1995). In the next iteration, active is a singleton set containing the state pop-ed out the top of the priority queue.
IDL-NGLM-BEAM is obtained by modifying line 8 to obtain:
IDL-NGLM-BEAM(G,LM,beam) | |
1 active {[vs^{G}]} | |
2 flag 1 | |
3 while flag | |
4 do unfold UNFOLDIDLG(active, G) | |
5 EVALUATENGLM(unfold, LM) | |
6 if FINALIDLG(unfold, G) | |
7 then flag 0 | |
8 active BEAMSTATES(unfold, beam) | |
9 return active | |
In this embodiment, the unfolding is controlled using a probabilistic beam beam, which, via the BEAMSTATES function, selects as active states only the states in unfold reachable with a probability higher or equal to the current maximum probability times the probability beam beam.
The IDL representation is well suited for computing accurate admissible heuristics under language models. These heuristics are needed by the IDL-NGLM-A*algorithm, and are also employed for pruning by the IDL-NGLM-BEAM algorithm.
For each state S in a weighted finite-state acceptor W corresponding to an IDL-graph G, one can efficiently extract from G—without further unfolding the set ‘of all edge labels that can be used to reach the final states of W. This set of labels, denoted FE s/all, is an overestimation of the set of future events reachable from S, because the labels under the V operators are all considered. From FE S/all and the n-l labels (when using an n-gram language model) recorded in state S the set of label sequences of length n-l is obtained. This set, denoted FCEs, is an (over)estimated set of possible future conditioning events for state S, which has the most cost-efficient future conditioning events for state S. Using FCEs, one needs to extract from FE: the set of most cost-efficient future events from under each v operator. This set, denoted FCEs, is used to arrive at an admissible heuristic for state S under a model LM, using Equation 2a:
If h*(S) is the true future cost for state S, this guarantees that h(S)<=h*(S)from the way FEs and FCEs are constructed. Note that, as it often happens with admissible heuristics, h(S) can come arbitrarily close to h*(S), by computing increasingly better approximations FCEs of FCE*s. Such approximations, however, require increasingly advanced unfoldings of the IDL-graph G (a complete unfolding of G for state S gives FCEs=FCE*s, and consequently h(S)=h*(S)). It follows that arbitrarily accurate admissible heuristics exist for IDL-expressions, but computing them only requires finding a balance between the time and space requirements for computing better heuristics and the speed-up obtained by using them in the search algorithms.
Formal Properties of IDL-NGLM Algorithms
The following theorem states the correctness of our algorithms, in the sense that they find the maximum path encoded by an IDL-graph under an gram language model.
Theorem 1 Let K be an IDL-expression, G(Π) its IDL-graph, and W(T) its wFSA under an n-gram language model LM. Algorithms NGLM-BFS and IDL-NGLM-A* find the path of maximum probability under LM. Algorithm IDL-NGLM-BEAM finds the path of maximum probability under LM, if all states in W(T) along this path are selected by its BEAMSTATES function. The proof of the theorem follows directly from the correctness of the BFS and A*search, and from the condition imposed on the beam search.
The next theorem characterizes the run-time complexity of these algorithms, in terms of an input IDL expression Π and its corresponding IDL-graph G(Π) complexity. There are three factors that linearly influence the run-time complexity of these algorithms: a is the maximum number of nodes in G(Π) needed to represent a state in A(Π)−a depends solely on Π; w is the maximum number of nodes in G(Π) needed represent a state in W(Π)−w depends on Π and n, the length of the context used by the n-gram language model; and K is the number of states of W (Π)−K also depends on Π and n. Of these three factors, K is by far the predominant one, and we simply call K the complexity of an IDL-expression.
Theorem 2 Let π be an IDL-expression, G(π) its | |
IDL-graph, A(π) its FSA, and W(π) its wFSA | |
under an n-gram language model. Let V[A(π)] | |
be the set of states of A(π), and V[W(π)] | |
the set of states of W(π). Let also a = | |
max_{c∈V[A(π)]}|c|,w = max_{c∈V[W(π)]}|c|, and | |
K = |V[W(π)]|. Algorithms IDL-NGLM-BFS | |
and IDL-NGLM-BEAM have run-time complexity | |
O(awK). Algorithm IDL-NGLM-A* has run-time | |
complexity O(awK log K). | |
These techniques are linear in the complexity of the input IDL-expression (with an additional log factor in the case of A* search due to priority queue management) which may be a significant advantage. Depending on the input IDL-expression, the task addressed can vary in complexity from linear to exponential. That is, for the intersection of an IDL-expression.
Intersection of WIDL-expressions with N-gram Language Models can thus be carried out as described above. The techniques for intersecting IDL-expressions with n-gram language models can be used almost without modification for intersecting WIDL-expressions with n-gram language models. These algorithms operate over IDL-graphs, and both IDL-graphs and WIDLgraphs are the same kind of graphs, over which algorithms operate in the same manner. The only modification to the algorithms presented in (Soricut and Marcu, 2005) consists on replacing line 5 in each of the IDL-NGLM-BFS, IDL-NGLM-A*, and IDL-NGLM-BEAM algorithms with a function evaluating the set of unfold state s under both the language model distributions and the input WIDL-expression distributions: EVALUATEWIDL-NGLM (unfold, G, LM).
The Result is a New Family of Algorithms:
WIDL-NGLM-BFS, WIDL-NGLM-A*, and WIDL-NGLM-BEAM, which perform intersection between the ranked finite languages of WIDL-expressions and n-gram language models. Similar with the IDL representation, accurate admissible heuristics can be computed for WIDLexpressions as well. In fact, Equation (2) above computes also an admissible heuristic for WIDL-expressions as well, as it considers the WIDL costs to be 0 (an underestimation) for each state in the weighted finite state acceptor. More accurate admissible heuristics, however, can be computed, and the result is lower degree of unfolding of the WIDL-expressions and faster search.
The intersection of IDL and WIDL-expressions with syntax-based language models is performed using a similar approach. A (W)IDL-expression is first mapped into a (W)IDL-graph, which is unfolded on-the-fly into a weighted finite state acceptor. Admissible heuristics are used to compute cost estimates for each state created in the finite state acceptor. The cost estimates are used in deciding what parts of the (W)IDL-graph are unfolded, and in what order. To arrive at the cost estimates, the syntax-based language model is used in evaluating prefix syntax trees, corresponding to the prefix strings unfolded from the (W)IDL-graph, as well as in evaluating the future costs, by estimating the cost of the most probable syntax tree which is consistent with the prefix tree already created. The resulting admissible heuristics are used to create a family of algorithms for intersecting (W)IDL-expressions with syntax-based language models, such as A* search algorithms and probabilistic beam search algorithms. Similar with the algorithms for n-gram language model intersection, the only difference between these types of algorithms consists in the selection of the finite state acceptor states which are used for the next round of on-the-fly unfolding of the WIDL-graph.)
Although only a few embodiments have been disclosed in detail above, other modifications are possible, and this disclosure is intended to cover all such modifications, and most particularly, any modification which might be predictable to a person having ordinary skill in the art. For example, other operators, besides those described herein, may be used. Also, while the above describes certain techniques of mapping to and intersecting with the natural language model, other techniques of doing this are contemplated. Moreover, while the above has described, and is predominantly directed to, weighted IDL, it should be understood that many of the techniques described here and be used with other kinds of IDL.
Also, only those claims which use the words “means for” are intended to be interpreted under 35 USC 112, sixth paragraph. Moreover, no limitations from the specification are intended to be read into any claims, unless those limitations are expressly included in the claims.
Cited Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|
US4502128 | Jun 1, 1982 | Feb 26, 1985 | Hitachi, Ltd. | Translation between natural languages |
US4599691 | May 13, 1983 | Jul 8, 1986 | Kokusai Denshin Denwa Co., Ltd. | Tree transformation system in machine translation system |
US4787038 | Mar 24, 1986 | Nov 22, 1988 | Kabushiki Kaisha Toshiba | Machine translation system |
US4814987 | May 19, 1986 | Mar 21, 1989 | Sharp Kabushiki Kaisha | Translation system |
US4942526 | Oct 24, 1986 | Jul 17, 1990 | Hitachi, Ltd. | Method and system for generating lexicon of cooccurrence relations in natural language |
US5146405 | Feb 5, 1988 | Sep 8, 1992 | At&T Bell Laboratories | Methods for part-of-speech determination and usage |
US5181163 | Jul 31, 1989 | Jan 19, 1993 | Hitachi, Ltd. | Method and apparatus for generating and/or updating cooccurrence relation dictionary |
US5212730 | Jul 1, 1991 | May 18, 1993 | Texas Instruments Incorporated | Voice recognition of proper names using text-derived recognition models |
US5267156 | Dec 5, 1991 | Nov 30, 1993 | International Business Machines Corporation | Method for constructing a knowledge base, knowledge base system, machine translation method and system therefor |
US5311429 | May 15, 1990 | May 10, 1994 | Hitachi, Ltd. | Maintenance support method and apparatus for natural language processing system |
US5432948 | Apr 26, 1993 | Jul 11, 1995 | Taligent, Inc. | Object-oriented rule-based text input transliteration system |
US5477451 | Jul 25, 1991 | Dec 19, 1995 | International Business Machines Corp. | Method and system for natural language translation |
US5510981 | Oct 28, 1993 | Apr 23, 1996 | International Business Machines Corporation | Language translation apparatus and method using context-based translation models |
US5644774 | Apr 25, 1995 | Jul 1, 1997 | Sharp Kabushiki Kaisha | Machine translation system having idiom processing function |
US5696980 | Apr 27, 1993 | Dec 9, 1997 | Sharp Kabushiki Kaisha | Machine translation system utilizing bilingual equivalence statements |
US5724593 | Jan 24, 1997 | Mar 3, 1998 | International Language Engineering Corp. | Machine assisted translation tools |
US5761631 | Jul 12, 1995 | Jun 2, 1998 | International Business Machines Corporation | Parsing method and system for natural language processing |
US5781884 | Nov 22, 1996 | Jul 14, 1998 | Lucent Technologies, Inc. | Grapheme-to-phoneme conversion of digit strings using weighted finite state transducers to apply grammar to powers of a number basis |
US5805832 | Jun 2, 1995 | Sep 8, 1998 | International Business Machines Corporation | System for parametric text to text language translation |
US5848385 | Nov 27, 1995 | Dec 8, 1998 | Sharp Kabushiki Kaisha | Machine translation system using well formed substructures |
US5867811 | Jun 17, 1994 | Feb 2, 1999 | Canon Research Centre Europe Ltd. | Method, an apparatus, a system, a storage device, and a computer readable medium using a bilingual database including aligned corpora |
US5870706 | Apr 10, 1996 | Feb 9, 1999 | Lucent Technologies, Inc. | Method and apparatus for an improved language recognition system |
US5903858 | Jun 28, 1996 | May 11, 1999 | Saraki; Masashi | Translation machine for editing a original text by rewriting the same and translating the rewrote one |
US5987404 | Jan 29, 1996 | Nov 16, 1999 | International Business Machines Corporation | Statistical natural language understanding using hidden clumpings |
US5991710 | May 20, 1997 | Nov 23, 1999 | International Business Machines Corporation | Statistical translation system with features based on phrases or groups of words |
US6031984 | Mar 9, 1998 | Feb 29, 2000 | I2 Technologies, Inc. | Method and apparatus for optimizing constraint models |
US6032111 | Jun 23, 1997 | Feb 29, 2000 | At&T Corp. | Method and apparatus for compiling context-dependent rewrite rules and input strings |
US6092034 | Jul 27, 1998 | Jul 18, 2000 | International Business Machines Corporation | Statistical translation system and method for fast sense disambiguation and translation of large corpora using fertility models and sense models |
US6119077 | Feb 6, 1997 | Sep 12, 2000 | Sharp Kasbushiki Kaisha | Translation machine with format control |
US6131082 | Dec 1, 1997 | Oct 10, 2000 | Int'l.Com, Inc. | Machine assisted translation tools utilizing an inverted index and list of letter n-grams |
US6182014 | Nov 20, 1998 | Jan 30, 2001 | Schlumberger Technology Corporation | Method and system for optimizing logistical operations in land seismic surveys |
US6205456 | Jan 13, 1998 | Mar 20, 2001 | Fujitsu Limited | Summarization apparatus and method |
US6223150 | Jan 29, 1999 | Apr 24, 2001 | Sony Corporation | Method and apparatus for parsing in a spoken language translation system |
US6236958 | May 15, 1998 | May 22, 2001 | International Business Machines Corporation | Method and system for extracting pairs of multilingual terminology from an aligned multilingual text |
US6278967 | Apr 23, 1996 | Aug 21, 2001 | Logovista Corporation | Automated system for generating natural language translations that are domain-specific, grammar rule-based, and/or based on part-of-speech analysis |
US6285978 | Sep 24, 1998 | Sep 4, 2001 | International Business Machines Corporation | System and method for estimating accuracy of an automatic natural language translation |
US6289302 | Aug 10, 1999 | Sep 11, 2001 | Matsushita Electric Industrial Co., Ltd. | Chinese generation apparatus for machine translation to convert a dependency structure of a Chinese sentence into a Chinese sentence |
US6304841 | Jul 29, 1997 | Oct 16, 2001 | International Business Machines Corporation | Automatic construction of conditional exponential models from elementary features |
US6311152 | Apr 8, 1999 | Oct 30, 2001 | Kent Ridge Digital Labs | System for chinese tokenization and named entity recognition |
US6360196 | May 18, 1999 | Mar 19, 2002 | Sharp Kabushiki Kaisha | Method of and apparatus for retrieving information and storage medium |
US6389387 | May 27, 1999 | May 14, 2002 | Sharp Kabushiki Kaisha | Method and apparatus for multi-language indexing |
US6393388 | Jun 5, 2000 | May 21, 2002 | Sony Corporation | Example-based translation method and system employing multi-stage syntax dividing |
US6393389 | Sep 23, 1999 | May 21, 2002 | Xerox Corporation | Using ranked translation choices to obtain sequences indicating meaning of multi-token expressions |
US6415250 | Jun 18, 1997 | Jul 2, 2002 | Novell, Inc. | System and method for identifying language using morphologically-based techniques |
US6460015 | Dec 15, 1998 | Oct 1, 2002 | International Business Machines Corporation | Method, system and computer program product for automatic character transliteration in a text string object |
US6502064 | Aug 31, 1998 | Dec 31, 2002 | International Business Machines Corporation | Compression method, method for compressing entry word index data for a dictionary, and machine translation system |
US6782356 | Oct 3, 2000 | Aug 24, 2004 | Hewlett-Packard Development Company, L.P. | Hierarchical language chunking translation table |
US6810374 | Jul 23, 2001 | Oct 26, 2004 | Pilwon Kang | Korean romanization system |
US6904402 | Jun 30, 2000 | Jun 7, 2005 | Microsoft Corporation | System and iterative method for lexicon, segmentation and language model joint optimization |
US7107215 | Apr 16, 2001 | Sep 12, 2006 | Sakhr Software Company | Determining a compact model to transcribe the arabic language acoustically in a well defined basic phonetic study |
US7113903 | Jan 30, 2002 | Sep 26, 2006 | At&T Corp. | Method and apparatus for providing stochastic finite-state machine translation |
US20020188438 | May 31, 2002 | Dec 12, 2002 | Kevin Knight | Integer programming decoder for machine translation |
US20020198701 | Jun 17, 2002 | Dec 26, 2002 | Moore Robert C. | Statistical method and apparatus for learning translation relationships among words |
US20040030551 | Mar 27, 2003 | Feb 12, 2004 | Daniel Marcu | Phrase to phrase joint probability model for statistical machine translation |
EP0469884A2 | Jul 31, 1991 | Feb 5, 1992 | Canon Kabushiki Kaisha | Sentence generating system |
EP0715265A2 | Nov 28, 1995 | Jun 5, 1996 | Sharp Kabushiki Kaisha | Machine translation system |
EP0933712A2 | Jan 14, 1999 | Aug 4, 1999 | Xerox Corporation | Method and system for generating document summaries with navigation information |
JPH1011447A | Title not available | |||
JPH07244666A | Title not available | |||
JPH11272672A | Title not available |
Reference | ||
---|---|---|
1 | Abney, Stephen, "Parsing by Chunks," 1991, Principle-Based Parsing: Computation and Psycholinguistics, vol. 44, pp. 257-279. | |
2 | Al-Onaizan et al., "Statistical Machine Translation," 1999, JHU Summer Tech Workshop, Final Report, pp. 1-42. | |
3 | Al-Onaizan et al., "Translating with Scarce Resources," 2000, 17th National Conference of the American Association for Artificial Intelligence, Austin, TX, pp. 672-678. | |
4 | Al-Onaizan, Y. and Knight, K., "Named Entity Translation: Extended Abstract" 2002, Proceedings of HLT-02, San Diego, CA. | |
5 | Al-Onaizan, Y. and Knight, K., "Translating Named Entities Using Monolingual and Bilingual Resources," 2002, Proc. of the 40th Annual Meeting of the ACL,pp. 400-408. | |
6 | Alshawi et al., "Learning Dependency Translation Models as Collections of Finite-State Head Transducers," 2000, Computational Linguistics, vol. 26, pp. 45-60. | |
7 | Arbabi et al., "Algorithms for Arabic name transliteration," Mar. 1994, IBM Journal of Research and Development, vol. 38, Issue 2, pp. 183-194. | |
8 | Bangalore, S. and Rambow, O., "Corpus-Based Lexical Choice in Natural Language Generation," 2000, Proc. of the 38th Annual ACL, Hong Kong, pp. 464-471. | |
9 | Bangalore, S. And Rambow, O., "Evaluation Metrics for Generation," 2000, Proc. of the 1st International Natural Language Generation Conf., vol. 14, p. 1-8. | |
10 | Bangalore, S. and Rambow, O., "Exploiting a Probabilistic Hierarchical Model for Generation," 2000, Proc. of 18th conf. on Computational Linguistics, vol. 1, pp. 42-48. | |
11 | Bangalore, S. and Rambow, O., "Using TAGs, a Tree Model, and a Language Model for Generation," May 2000, Workshop TAG+5, Paris. | |
12 | Barnett et al., "Knowledge and Natural Language Processing," Aug. 1990, Communications of the ACM, vol. 33, Issue 8, pp. 50-71. | |
13 | Baum, Leonard, "An Inequality and Associated Maximization Technique in Statistical Estimation for Probabilistic Functions of Markov Processes", 1972, Inequalities 3:1-8. | |
14 | Bikel et al., "An Algorithm that Learns What's in a Name," 1999, Machine Learning Journal Special Issue on Natural Language Learning, vol. 34, pp. 211-232. | |
15 | Brants, Thorsten, "TnT-A Statistical Part-of-Speech Tagger," 2000, Proc. of the 6th Applied Natural Language Processing Conference, Seattle. | |
16 | Brants, Thorsten, "TnT—A Statistical Part-of-Speech Tagger," 2000, Proc. of the 6th Applied Natural Language Processing Conference, Seattle. | |
17 | Brill, Eric. "Transformation-Based Error-Driven Learning and Natural Language Processing: A Case Study in Part of Speech Tagging", 1995, Computational Linguistics, vol. 21, No. 4, pp. 543-565. | |
18 | Brown et al., "A Statistical Approach to Machine Translation," Jun. 1990, Computational Linguistics, vol. 16, No. 2, pp. 79-85. | |
19 | Brown et al., "The Mathematics of Statistical Machine Translation: Parameter Estimation," 1993, Computational Linguistics, vol. 19, Issue 2, pp. 263-311. | |
20 | Brown et al., "Word-Sense Disambiguation Using Statistical Methods," 1991, Proc. of 29th Annual ACL, pp. 264-270. | |
21 | Brown, Ralf, "Automated Dictionary Extraction for "Knowledge-Free" Example-Based Translation," 1997, Proc. of 7th Int'l Conf. on Theoretical and Methodological Issues in MT, Santa Fe, NM, pp. 111-118. | |
22 | Callan et al., "TREC and TIPSTER Experiments with Inquery," 1994, Information Processing and Management, vol. 31, Issue 3, pp. 327-343. | |
23 | Carl, Michael. "A Constructivist Approach to Machine Translation," 1998, New Methods of Language Processing and Computational Natural Language Learning, pp. 247-256. | |
24 | Chen, K. and Chen, H., "Machine Translation: An Integrated Approach," 1995, Proc. of 6th Int'l Conf. on Theoretical and Methodological Issue in MT, pp. 287-294. | |
25 | Chinchor, Nancy, "MUC-7 Named Entity Task Definition," 1997, Version 3.5. | |
26 | Clarkson, P. and Rosenfeld, R., "Statistical Language Modeling Using the CMU-Cambridge Toolkit", 1997, Proc. ESCA Eurospeech, Rhodes, Greece, pp. 2707-2710. | |
27 | Cohen, Yossi, "Interpreter for FUF," (available at ftp://ftp.cs.bgu.ac.il/pub/people/elhadad/fuf-life.lf). | |
28 | Corston-Oliver, Simon, "Beyond String Matching and Cue Phrases: Improving Efficiency and Coverage in Discourse Analysis",1998, The AAAI Spring Symposium on Intelligent Text Summarization, pp. 9-15. | |
29 | Dagan, I. and Itai, A., "Word Sense Disambiguation Using a Second Language Monolingual Corpus", 1994, Computational Linguistics, vol. 20, No. 4, pp. 563-596. | |
30 | Dempster et al., "Maximum Likelihood from Incomplete Data via the EM Algorithm", 1977, Journal of the Royal Statistical Society, vol. 39, No. 1, pp. 1-38. | |
31 | Diab, M. and Finch, S., "A Statistical Word-Level Translation Model for Comparable Corpora," 2000, In Proc.of the Conference on ContentBased Multimedia Information Access (RIAO). | |
32 | Diab, Mona, "An Unsupervised Method for Multilingual Word Sense Tagging Using Parallel Corpora: A Preliminary Investigation", 2000, SIGLEX Workshop on Word Senses and Multi-Linguality, pp. 1-9. | |
33 | Elhadad et al., "Floating Constraints in Lexical Choice", 1996, ACL, 23(2): 195-239. | |
34 | Elhadad, M. and Robin, J., "An Overview of SURGE: a Reusable Comprehensive Syntactic Realization Component," 1996, Technical Report 96-03, Department of Mathematics and Computer Science, Ben Gurion University, Beer Sheva, Israel. | |
35 | Elhadad, M. and Robin, J., "Controlling Content Realization with Functional Unification Grammars", 1992, Aspects of Automated Natural Language Generation, Dale et al. (eds)., Springer Verlag, pp. 89-104. | |
36 | Elhadad, Michael, "FUF: the Universal Unifier User Manual Version 5.2", 1993, Department of Computer Science, Ben Gurion University, Beer Sheva, Israel. | |
37 | Elhadad, Michael, "Using Argumentation to Control Lexical Choice: A Functional Unification Implementation", 1992, Ph.D. Thesis, Graduate School of Arts and Sciences, Columbia University. | |
38 | Elhadad. M. and Robin, J., "SURGE: a Comprehensive Plug-in Syntactic Realization Component for Text Generation", 1999 (available at http://www.cs.bgu.ac.il/~elhadad/pub.html). | |
39 | Elhadad. M. and Robin, J., "SURGE: a Comprehensive Plug-in Syntactic Realization Component for Text Generation", 1999 (available at http://www.cs.bgu.ac.il/˜elhadad/pub.html). | |
40 | Fung, P. and Yee, L., "An IR Approach for Translating New Words from Nonparallel, Comparable Texts", 1998, 36th Annual Meeting of the ACL, 17th International Conference on Computational Linguistics, pp. 414-420. | |
41 | Fung, Pascale, "Compiling Bilingual Lexicon Entries From a Non-Parallel English-Chinese Corpus", 1995, Proc. of the Third Workshop on Very Large Corpora, Boston, MA, pp. 173-183. | |
42 | Gale, W. and Church, K., "A Program for Aligning Sentences in Bilingual Corpora," 1991, 29th Annual Meeting of the ACL, pp. 177-183. | |
43 | Germann et al., "Fast Decoding and Optimal Decoding for Machine Translation", 2001, Proc. of the 39th Annual Meeting of the ACL, Toulouse, France, pp. 228-235. | |
44 | Germann, Ulrich, "Building a Statistical Machine Translation System from Scratch: How Much Bang for the Buck Can We Expect?" Proc. of the Data-Driven MT Workshop of ACL-01, Toulouse, France, 2001. | |
45 | Grefenstette, Gregory, "The World Wide Web as a Resource for Example-Based Machine Translation Tasks", 1999, Translating and the Computer 21, Proc. of the 21st International Conf. on Translating and the Computer, London, UK, 12 pp. | |
46 | Hatzivassiloglou, V. et al., "Unification-Based Glossing", 1995, Proc. of the International Joint Conference on Artificial Intelligence, pp. 1382-1389. | |
47 | Ide, N. and Veronis, J., "Introduction to the Special Issue on Word Sense Disambiguation: The State of the Art", Mar. 1998, Computational Linguistics, vol. 24, Issue 1, pp. 2-40. | |
48 | Imamura, Kenji, "Hierarchical Phrase Alignment Harmonized with Parsing", 2001, in Proc. of NLPRS, Tokyo. | |
49 | Jelinek, F., "Fast Sequential Decoding Algorithm Using a Stack", Nov. 1969, IBM J. Res. Develop., vol. 13, No. 6, pp. 675-685. | |
50 | Jones, K. Sparck, "Experiments in Relevance Weighting of Search Terms", 1979, Information Processing & Management, vol. 15, Pergamon Press Ltd., UK, pp. 133-144. | |
51 | Knight et al., "Filling Knowledge Gaps in a Broad-Coverage Machine Translation System", 1995, Proc. of the14th International Joint Conference on Artificial Intelligence, Montreal, Canada, vol. 2, pp. 1390-1396. | |
52 | Knight et al., "Integrating Knowledge Bases and Statistics in MT," 1994, Proc. of the Conference of the Association for Machine Translation in the Americas. | |
53 | Knight et al., "Translation with Finite-State Devices," 1998, Proc. of the 3rd AMTA Conference, pp. 421-437. | |
54 | Knight, K. and Al-Onaizan, Y., "A Primer on Finite-State Software for Natural Language Processing", 1999 (available at http://www.isi.edu/licensed-sw/carmel). | |
55 | Knight, K. and Chander, I., "Automated Postediting of Documents,"1994, Proc. of the 12th Conference on Artificial Intelligence, pp. 779-784. | |
56 | Knight, K. and Graehl, J., "Machine Transliteration", 1997, Proc. of the ACL-97, Madrid, Spain. | |
57 | Knight, K. and Hatzivassiloglou, V., "Two-Level, Many-Paths Generation," 1995, Proc. of the 33rd Annual Conference of the ACL, pp. 252-260. | |
58 | Knight, K. and Luk, S., "Building a Large-Scale Knowledge Base for Machine Translation," 1994, Proc. of the 12th Conference on Artificial Intelligence, pp. 773-778. | |
59 | Knight, K. and Marcu, D., "Statistics-Based Summarization-Step One: Sentence Compression," 2000, American Association for Artificial Intelligence Conference, pp. 703-710. | |
60 | Knight, K. and Marcu, D., "Statistics-Based Summarization—Step One: Sentence Compression," 2000, American Association for Artificial Intelligence Conference, pp. 703-710. | |
61 | Knight, K. and Yamada, K., "A Computational Approach to Deciphering Unknown Scripts," 1999, Proc. of the ACL Workshop on Unsupervised Learning in Natural Language Processing. | |
62 | Knight, K. et al., "Machine Transliteration of Names in Arabic Text," 2002, Proc. of the ACL Workshop on Computational Approaches to Semitic Languages. | |
63 | Knight, Kevin, "A Statistical MT Tutorial Workbook," 1999, JHU Summer Workshop (available at http://www.isi.edu/natural-language/mt/wkbk.rtf). | |
64 | Knight, Kevin, "Automating Knowledge Acquisition for Machine Translation," 1997, AI Magazine 18(4). | |
65 | Knight, Kevin, "Connectionist Ideas and Algorithms," Nov. 1990, Communications of the ACM, vol. 33, No. 11, pp. 59-74. | |
66 | Knight, Kevin, "Decoding Complexity in Word-Replacement Translation Models", 1999, Computational Linguistics, 25(4). | |
67 | Knight, Kevin, "Integrating Knowledge Acquisition and Language Acquisition," May 1992, Journal of Applied Intelligence, vol. 1, No. 4. | |
68 | Knight, Kevin, "Learning Word Meanings by Instruction,"1996, Proc. of the National Conference on Artificial Intelligence, vol. 1, pp. 447-454. | |
69 | Knight, Kevin, "Unification: A Multidisciplinary Survey," 1989, ACM Computing Surveys, vol. 21, No. 1. | |
70 | Koehn, P. and Knight, K., "ChunkMT: Statistical Machine Translation with Richer Linguistic Knowledge," Apr. 2002, Information Sciences Institution. | |
71 | Koehn, P. and Knight, K., "Estimating Word Translation Probabilities from Unrelated Monolingual Corpora Using the EM Algorithm," 2000, Proc. of the 17th meeting of the AAAI. | |
72 | Koehn, P. and Knight, K., "Knowledge Sources for Word-Level Translation Models," 2001, Conference on Empirical Methods in Natural Language Processing. | |
73 | Kumar, R. and Li, H., "Integer Programming Approach to Printed Circuit Board Assembly Time Optimization," 1995, IEEE Transactions on Components, Packaging, and Manufacturing, Part B: Advance Packaging, vol. 18, No. 4, pp. 720-727. | |
74 | Kurohashi, S. and Nagao, M., "Automatic Detection of Discourse Structure by Checking Surface Information in Sentences," 1994, Proc. of COL-LING '94, vol. 2, pp. 1123-1127. | |
75 | Langkilde, I. and Knight, K., "Generation that Exploits Corpus-Based Statistical Knowledge," 1998, Proc. of the COLING-ACL, pp. 704-710. | |
76 | Langkilde, I. and Knight, K., "The Practical Value of N-Grams in Generation," 1998, Proc. of the 9th International Natural Language Generation Workshop, p. 248-255. | |
77 | Langkilde, Irene, "Forest-Based Statistical Sentence Generation," 2000, Proc. of the 1st Conference on North American chapter of the ACL, Seattle, WA, pp. 170-177. | |
78 | Langkilde-Geary, Irene, "A Foundation for General-Purpose Natural Language Generation: Sentence Realization Using Probabilistic Models of Language," 2002, Ph.D. Thesis, Faculty of the Graduate School, University of Southern California. | |
79 | Langkilde-Geary, Irene, "An Empirical Verification of Coverage and Correctness for a General-Purpose Sentence Generator," 1998, Proc. 2nd Int'l Natural Language Generation Conference. | |
80 | Mann, G. and Yarowsky, D., "Multipath Translation Lexicon Induction via Bridge Languages," 2001, Proc. of the 2nd Conference of the North American Chapter of the ACL, Pittsburgh, PA, pp. 151-158. | |
81 | Manning, C. and Schutze, H., "Foundations of Statistical Natural Language Processing," 2000, The MIT Press, Cambridge, MA [redacted]. | |
82 | Marcu, D. and Wong, W., "A Phrase-Based, Joint Probability Model for Statistical Machine Translation," 2002, Proc. of ACL-2 conference on Empirical Methods in Natural Language Processing, vol. 10, pp. 133-139. | |
83 | Marcu, Daniel, "Building Up Rhetorical Structure Trees," 1996, Proc. of the National Conference on Artificial Intelligence and Innovative Applications of Artificial Intelligence Conference, vol. 2, pp. 1069-1074. | |
84 | Marcu, Daniel, "Discourse trees are good indicators of importance in text," 1999, Advances in Automatic Text Summarization, The MIT Press, Cambridge, MA. | |
85 | Marcu, Daniel, "Instructions for Manually Annotating the Discourse Structures of Texts," 1999, Discourse Annotation, pp. 1-49. | |
86 | Marcu, Daniel, "The Rhetorical Parsing of Natural Language Texts," 1997, Proceedings of ACL/EACL '97, pp. 96-103. | |
87 | Marcu, Daniel, "The Rhetorical Parsing, Summarization, and Generation of Natural Language Texts," 1997, Ph.D. Thesis, Graduate Department of Computer Science, University of Toronto. | |
88 | Marcu, Daniel, "Towards a Unified Approach to Memory- and Statistical-Based Machine Translation," 2001, Proc. of the 39th Annual Meeting of the ACL, pp. 378-385. | |
89 | Melamed, I. Dan, "A Word-to-Word Model of Translational Equivalence," 1997, Proc. of the 35th Annual Meeting of the ACL, Madrid, Spain, pp. 490-497. | |
90 | Melamed, I. Dan, "Automatic Evaluation and Uniform Filter Cascades for Inducing N-Best Translation Lexicons," 1995, Proc. of the 3rd Workshop on Very Large Corpora, Boston, MA, pp. 184-198. | |
91 | Melamed, I. Dan, "Empirical Methods for Exploiting Parallel Texts," 2001, MIT Press, Cambridge, MA [table of contents]. | |
92 | Meng et al., "Generating Phonetic Cognates to Handle Named Entities in English-Chinese Cross-Language Spoken Document Retrieval," 2001, IEEE Workshop on Automatic Speech Recognition and Understanding, pp. 311-314. | |
93 | Miike et al., "A full-text retrieval system with a dynamic abstract generation function," 1994, Proceedings of SI-GIR '94, pp. 152-161. | |
94 | Mikheev et al., "Named Entity Recognition without Gazeteers," 1999, Proc. of European Chapter of the ACL, Bergen, Norway, pp. 1-8. | |
95 | Mohri, M. and Riley, M., "An Efficient Algorithm for the N-Best-Strings Problem," 2002, Proc. of the 7th Int. Conf. on Spoken Language Processing (ICSLP'02), Denver, CO, pp. 1313-1316. | |
96 | Monasson et al., "Determining computational complexity from characteristic ‘phase transitions’," Jul. 1999, Nature Magazine, vol. 400, pp. 133-137. | |
97 | Monasson et al., "Determining computational complexity from characteristic 'phase transitions'," Jul. 1999, Nature Magazine, vol. 400, pp. 133-137. | |
98 | Mooney, Raymond, "Comparative Experiments on Disambiguating Word Senses: An Illustration of the Role of Bias in Machine Learning," 1996, Proc. of the Conference on Empirical Methods in Natural Language Processing, pp. 82-91. | |
99 | * | Nederhof et al.,IDL-Expressions: A formalism for Representing and Parsing Finite Languages in Natural Languages Processing, Mar. 2004; Journal of Artificial Intelligence Research 21; pp. 287-317. |
100 | Nederhof, M. and Satta, G., "IDL-Expressions: A Formalism for Representing and Parsing Finite Languages in Natural Language Processing," 2004, Journal of Artificial Intelligence Research, vol. 21, pp. 281-287. | |
101 | Niessen,S. and Ney, H, "Toward hierarchical models for statistical machine translation of inflected languages," 2001, Data-Driven Machine Translation Workshop, Toulouse, France, pp. 47-54. | |
102 | Och et al., "Improved Alignment Models for Statistical Machine Translation," 1999, Proc. of the Joint Conf. of Empirical Methods in Natural Language Processing and Very Large Corpora, pp. 20-28. | |
103 | Och, F. and Ney, H, "Improved Statistical Alignment Models," 2000, 38th Annual Meeting of the ACL, Hong Kong, pp. 440-447. | |
104 | Och, F. and Ney, H., "Discriminative Training and Maximum Entropy Models for Statistical Machine Translation," 2002, Proc. of the 40th Annual Meeting of the ACL, Philadelphia, PA, pp. 295-302. | |
105 | Papineni et al., "Bleu: a Method for Automatic Evaluation of Machine Translation," 2001, IBM Research Report, RC22176(WO102-022). | |
106 | Pla et al., "Tagging and Chunking with Bigrams," 2000, Proc. of the 18th Conference on Computational Linguistics, vol. 2, pp. 614-620. | |
107 | Rapp, Reinhard, "Identifying Word Translations in Non-Parallel Texts," 1995, 33rd Annual Meeting of the ACL, pp. 320-322. | |
108 | Rapp, Reinhard, Automatic Identification of Word Translations from Unrelated English and German Corpora, 1999, 37th Annual Meeting of the ACL, pp. 519-526. | |
109 | Resnik, P. and Smith, A., "The Web as a Parallel Corpus," Sep. 2003, Computational Linguistics, Special Issue on Web as Corpus, vol. 29, Issue 3, pp. 349-380. | |
110 | Resnik, P. and Yarowsky, D., "A Perspective on Word Sense Disambiguation Methods and Their Evaluation," 1997, Proceedings of SIGLEX '97, Washington, DC, pp. 79-86. | |
111 | Resnik, Philip, "Mining the Web for Bilingual Text," 1999, 37th Annual Meeting of the ACL, College Park, MD, pp. 527-534. | |
112 | Rich, E. and Knight, K., "Artificial Intelligence, Second Edition," 1991, McGraw-Hill Book Company [redacted]. | |
113 | Richard et al., "Visiting the Traveling Salesman Problem with Petri nets and application in the glass industry," Feb. 1996, IEEE Emerging Technologies and Factory Automation, pp. 238-242. | |
114 | Robin, Jacques, "Revision-Based Generation of Natural Language Summaries Providing Historical Background: Corpus-Based Analysis, Design Implementation and Evaluation," 1994, Ph.D. Thesis, Columbia University, New York. | |
115 | Russell, S. and Norvig, P., "Artificial Intelligence: A Modern Approach," 1995, Prentice-Hall, Inc., New Jersey [redacted—table of contents]. | |
116 | Sang, E. and Buchholz, S., "Introduction to the CoNLL-2000 Shared Task: Chunking," 20002, Proc. of CoNLL-2000 and LLL-2000, Lisbon, Portugal, pp. 127-132. | |
117 | Schmid, H., and Walde, S., "Robust German Noun Chunking With a Probabilistic Context-Free Grammar," 2000, Proc. of the 18th Conference on Computational Linguistics, vol. 2, pp. 726-732. | |
118 | Schutze, Hinrich, "Automatic Word Sense Discrimination," 1998, Computational Linguistics, Special Issue on Word Sense Disambiguation, vol. 24, Issue 1, pp. 97-123. | |
119 | Selman et al., "A New Method for Solving Hard Satisfiability Problems," 1992, Proc. of the 10th National Conference on Artificial Intelligence, San Jose, CA, pp. 440-446. | |
120 | Shapiro, Stuart (ed.), "Encyclopedia of Artificial Intelligence, 2nd edition", vol. 2, 1992, John Wiley & Sons Inc; "Unification" article, K. Knight, pp. 1630-1637. | |
121 | Sobashima et al., "A Bidirectional Transfer-Driven Machine Translation System for Spoken Dialogues," 1994, Proc. of 15th Conference on Computational Linguistics, vol. 1, pp. 64-68. | |
122 | Soricut et al., "Using a large monolingual corpus to improve translation accuracy," 2002, Lecture Notes in Computer Science, vol. 2499, Proc. of the 5th Conference of the Association for Machine Translation in the Americas on Machine Translation: From Research to Real Users, pp. 155-164. | |
123 | Stalls, B. and Knight, K., "Translating Names and Technical Terms in Arabic Text," 1998, Proc. of the COLING/ACL Workkshop on Computational Approaches to Semitic Language. | |
124 | Sumita et al., "A Discourse Structure Analyzer for Japanese Text," 1992, Proc. of the International Conference on FIfth Generation Computer Systems, vol. 2, pp. 1133-1140. | |
125 | Sun et al., "Chinese Named Entity Identification Using Class-based Language Model," 2002, Proc. of 19th International Conference on Computational Linguistics, Taipei, Taiwan, vol. 1, pp. 1-7. | |
126 | Taylor et al., "The Penn Treebank: An Overview," in A. Abeill (ed.), Treebanks: Building and Using Parsed Corpora, 2003, pp. 5-22. | |
127 | Tiedemann, Jorg, "Automatic Construction of Weighted String Similarity Measures," 1999, In Proceedings of the Joint SIGDAT Conference on Emperical Methods in Natural Language Processing and Very Large Corpora. | |
128 | Tillman, C. and Xia, F., "A Phrase-Based Unigram Model for Statistical Machine Translation," 2003, Proc. of the North American Chapter of the ACL on Human Language Technology, vol. 2, pp. 106-108. | |
129 | Tillmann et al., "A DP based Search Using Monotone Alignments in Statistical Translation," 1997, Proc. of the Annual Meeting of the ACL, pp. 366-372. | |
130 | Ueffing et al., "Generation of Word Graphs in Statistical Machine Translation," 2002, Proc. of Empirical Methods in Natural Language Processing (EMNLP), pp. 156-163. | |
131 | Veale, T. and Way, A., "Gaijin: A Bootstrapping, Template-Driven Approach to Example-Based MT," 1997, Proc. of New Methods in Natural Language Processing (NEMPLP97), Sofia, Bulgaria. | |
132 | Vogel et al., "The CMU Statistical Machine Translation System," 2003, Machine Translation Summit IX, New Orleans, LA. | |
133 | Vogel et al., "The Statistical Translation Module in the Verbmobil System," 2000, Workshop on Multi-Lingual Speech Communication, pp. 69-74. | |
134 | Vogel, S. and Ney, H., "Construction of a Hierarchical Translation Memory," 2000, Proc. of Cooling 2000, Saarbrucken, Germany, pp. 1131-1135. | |
135 | Wang, Y. and Waibel, A., "Decoding Algorithm in Statistical Machine Translation," 1996, Proc. of the 35th Annual Meeting of the ACL, pp. 366-372. | |
136 | Wang, Ye-Yi, "Grammar Interference and Statistical Machine Translation," 1998, Ph.D Thesis, Carnegie Mellon University, Pittsburgh, PA. | |
137 | Watanbe et al., "Statistical Machine Translation Based on Hierarchical Phrase Alignment," 2002, 9th International Conference on Theoretical and Methodological Issues in Machin Translation (TMI-2002), Keihanna, Japan, pp. 188-198. | |
138 | Witbrock, M. and Mittal, V., "Ultra-Summarization: A Statistical Approach to Generating Highly Condensed Non-Extractive Summaries," 1999, Proc. of SIGIR '99, 22nd International Conference on Research and Development in Information Retrieval, Berkeley, CA, pp. 315-316. | |
139 | Wu, Dekai, "A Polynomial-Time Algorithm for Statistical Machine Translation," 1996, Proc. of 34th Annual Meeting of the ACL, pp. 152-158. | |
140 | Wu, Dekai, "Stochastic Inversion Transduction Grammars and Bilingual Parsing of Parallel Corpora," 1997, Computational Linguistics, vol. 23, Issue 3, pp. 377-403. | |
141 | Yamada, K. and Knight, K. "A Syntax-based Statistical Translation Model," 2001, Proc. of the 39th Annual Meeting of the ACL, pp. 523-530. | |
142 | Yamada, K. and Knight, K., "A Decoder for Syntax-based Statistical MT," 2001, Proceedings of the 40th Annual Meeting of the ACL, pp. 303-310. | |
143 | Yamamoto et al., "A Comparative Study on Translation Units for Bilingual Lexicon Extraction," 2001, Japan Academic Association for Copyright Clearance, Tokyo, Japan. | |
144 | Yarowsky, David, "Unsupervised Word Sense Disambiguation Rivaling Supervised Methods," 1995, 33rd Annual Meeting of the ACL, pp. 189-196. |
Citing Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|
US8214196 | Jul 3, 2002 | Jul 3, 2012 | University Of Southern California | Syntax-based statistical translation model |
US8234106 | Oct 8, 2009 | Jul 31, 2012 | University Of Southern California | Building a translation lexicon from comparable, non-parallel corpora |
US8296127 | Mar 22, 2005 | Oct 23, 2012 | University Of Southern California | Discovery of parallel text portions in comparable collections of corpora and training using comparable texts |
US8380486 | Oct 1, 2009 | Feb 19, 2013 | Language Weaver, Inc. | Providing machine-generated translations and corresponding trust levels |
US8433556 | Nov 2, 2006 | Apr 30, 2013 | University Of Southern California | Semi-supervised training for statistical word alignment |
US8468149 | Jan 26, 2007 | Jun 18, 2013 | Language Weaver, Inc. | Multi-lingual online community |
US8484016 * | May 28, 2010 | Jul 9, 2013 | Microsoft Corporation | Locating paraphrases through utilization of a multipartite graph |
US8543374 * | Aug 12, 2010 | Sep 24, 2013 | Xerox Corporation | Translation system combining hierarchical and phrase-based models |
US8548794 | Jul 2, 2004 | Oct 1, 2013 | University Of Southern California | Statistical noun phrase translation |
US8600728 | Oct 12, 2005 | Dec 3, 2013 | University Of Southern California | Training for a text-to-text application which uses string to tree conversion for training and decoding |
US8615389 | Mar 14, 2008 | Dec 24, 2013 | Language Weaver, Inc. | Generation and exploitation of an approximate language model |
US8666725 | Apr 15, 2005 | Mar 4, 2014 | University Of Southern California | Selection and use of nonstatistical translation components in a statistical machine translation framework |
US8676563 | Jun 21, 2010 | Mar 18, 2014 | Language Weaver, Inc. | Providing human-generated and machine-generated trusted translations |
US8694303 | Jun 15, 2011 | Apr 8, 2014 | Language Weaver, Inc. | Systems and methods for tuning parameters in statistical machine translation |
US8825466 | Jun 8, 2007 | Sep 2, 2014 | Language Weaver, Inc. | Modification of annotated bilingual segment pairs in syntax-based machine translation |
US8831928 | Apr 4, 2007 | Sep 9, 2014 | Language Weaver, Inc. | Customizable machine translation service |
US8886515 | Oct 19, 2011 | Nov 11, 2014 | Language Weaver, Inc. | Systems and methods for enhancing machine translation post edit review processes |
US8886517 | Jun 29, 2012 | Nov 11, 2014 | Language Weaver, Inc. | Trust scoring for language translation systems |
US8886518 | Aug 7, 2006 | Nov 11, 2014 | Language Weaver, Inc. | System and method for capitalizing machine translated text |
US8942973 | Mar 9, 2012 | Jan 27, 2015 | Language Weaver, Inc. | Content page URL translation |
US8943080 | Dec 5, 2006 | Jan 27, 2015 | University Of Southern California | Systems and methods for identifying parallel documents and sentence fragments in multilingual document collections |
US8977536 | Jun 3, 2008 | Mar 10, 2015 | University Of Southern California | Method and system for translating information with a higher probability of a correct translation |
US8990064 | Jul 28, 2009 | Mar 24, 2015 | Language Weaver, Inc. | Translating documents based on content |
US9063931 * | Feb 16, 2011 | Jun 23, 2015 | Ming-Yuan Wu | Multiple language translation system |
US9122674 | Dec 15, 2006 | Sep 1, 2015 | Language Weaver, Inc. | Use of annotations in statistical machine translation |
US9152622 | Nov 26, 2012 | Oct 6, 2015 | Language Weaver, Inc. | Personalized machine translation via online adaptation |
US9213694 | Oct 10, 2013 | Dec 15, 2015 | Language Weaver, Inc. | Efficient online domain adaptation |
US20110295589 * | May 28, 2010 | Dec 1, 2011 | Microsoft Corporation | Locating paraphrases through utilization of a multipartite graph |
US20120016661 * | Jul 19, 2010 | Jan 19, 2012 | Eyal Pinkas | System, method and device for intelligent textual conversation system |
US20120041753 * | Aug 12, 2010 | Feb 16, 2012 | Xerox Corporation | Translation system combining hierarchical and phrase-based models |
US20120209588 * | Feb 16, 2011 | Aug 16, 2012 | Ming-Yuan Wu | Multiple language translation system |
US20130054224 * | Aug 30, 2012 | Feb 28, 2013 | Dublin City University | Method and system for enhancing text alignment between a source language and a target language during statistical machine translation |
U.S. Classification | 704/9, 704/8, 704/7, 704/257 |
International Classification | G10L17/27 |
Cooperative Classification | G06F17/27 |
European Classification | G06F17/27 |
Date | Code | Event | Description |
---|---|---|---|
Oct 17, 2005 | AS | Assignment | Owner name: LANGUAGE WEAVER, INC., CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:SORICUTT, RADU;MARCU, DANIEL;REEL/FRAME:016897/0123 Effective date: 20050827 |
Dec 19, 2014 | FPAY | Fee payment | Year of fee payment: 4 |
Feb 16, 2016 | AS | Assignment | Owner name: SDL INC., MASSACHUSETTS Free format text: MERGER;ASSIGNOR:LANGUAGE WEAVER, INC.;REEL/FRAME:037745/0391 Effective date: 20151231 |